Metabolic engineering of yeast for xylose uptake and fermentation

Abstract

Xylose is the major pentose and the second most abundant sugar in lignocellulosic feedstocks. Therefore, the efficient utilization of xylose is required for the cost-effective bioconversion of lignocellulose. Rational and combinatorial metabolic engineering approaches coupled with transcriptomic and proteomic studies have been extensively exploited on the yeast Saccharomyces cerevisiae for improved xylose utilization. However, the resulting strains remain inapplicable for industrial ethanolic fermentation, although basic engineering approaches have been established and targets have been identified for further modification. In this study, we started with the rational genetic engineering of a laboratory S. cerevisiae strain to express the xylose metabolic pathways, including the xylose reductase/xylitol dehydrogenase (XR/XDH) and xylose isomerase (XI) pathways. The xylulokinase (XK) and the non-oxidative pentose phosphate pathway (PPP) were also overexpressed to facilitate pentose assimilation. The resulting strains, H131 -XYL 123 and H13 1-XYLA3 1, exhibited slow but significant aerobic growth (p.1ma=0.031+0.022 h~1 and 0.081+0.052 h-1, respectively), establishing a baseline for further advancement. These engineered strains mentioned above were then used to initiate a three-stage evolutionary engineering, through aerobic and oxygen-limited sequential batch cultivation followed by xylose-limited anaerobic chemostat cultivation. We aborted the development of the XR/XDH-based strain after the first stage of aerobic evolution, owing to the intrinsic cofactor imbalance of the two enzymes. In contrast, continuous improvement was observed during adaption of the XI-based H131-XYLA31, when strains were isolated from the evolved populations periodically and evaluated in terms of growth and fermentation. The final isolated strain, H131E8-XYLA31, displayed a significantly increased anaerobic growth rate (0.120+0.004 h-1) and xylose consumption rate (0.916 g-g h ) compared to its parent strain. Upon successful evolutionary engineering, the H131E8-XYLA31 was further modified by complementing the auxotrophic markers arg4 and leu2, resulting in H153E10-XYLA31 with greatly boosted aerobic growth. Moreover, adding the anaerobic growth factors ergosterol and Tween 80 to the medium enabled a maximum anaerobic growth rate of 0.199 h-1 and a specific xylose consumption rate of 1.647g-g-lh-1 in batch fermentation, 65% and 37% higher than those of the best reported xylosefermenting strain RWB 218, respectively. In chemostat cultivation, the strains exhibit specific performance more than 50% better than in the batch cultivation, implying the potential for further improvement of the strains in extended evolution. In order to identify the genotypes responsible for quick xylose utilization in the evolutionarily engineered strain, an inverse inverse metabolic engineering approach was applied to the evolved strain, based on functional complementation of the evolved H131E5-XYLA31 genomic library to an unevolved background strain. A highthroughput micro-fluidic screening method was used to screen the library, revealing the tandem duplication of XYLA in the H131E5-XYLA31 genome. The structure of XYLA integration, coupled with qPCR, DNA/RNA blotting, and enzyme activity assay results, suggests that the high expression level of XI is a major recombination event during the evolution and is necessary for efficient xylose assimilation. However, high XI expression does not necessarily imply the tandem multi-copy integration of XYLA into the genome. The XYLA expansion could originate from different initial constructions, subject to recombination/rearrangement, and produce different configurations. The effects of other metabolic engineering targets were also investigated, including other heterologous XIs, XK (XYL3), aldose reductase (GRE3), NADPdependent G3P dehydrogenase (gapN), and heterologous xylose transporters. Most results verified the previous hypothesis, leaving room for future strain improvements. In this study, we successfully applied rational and combinatorial metabolic engineering approaches for both constructing rapid xylose-fermenting strains and identifying novel genetic characteristics. We demonstrated the use of evolutionary engineering to achieve superior strains, as well as the inverse metabolic engineering approach based on micro-fluidic screening for both strain improvement and genotype identification. As such, the work of this thesis serves as a practical milestone in lignocellulose conversions. The constructed strains can be used as hosts for further advancement, and the metabolic engineering techniques have proven to be effective tools for future strain evolution.